If a option could be added to the kite-dataset CLI, so all columns get 'null', this would greatly help.
I'm currently using json-schema to create a avro schema, based on different json files from odata webservices. The prediction of what types should be, works ok;,
but being able to have all types [nullable] by default with an option, would greatly help.
So the importer wont fail as often.